Taxonomic notes on Cistanthe, Calandrinia, and Talinum (Portulacaceae)

17 Pages·1991·4.3 MB·

by  M A Hershkovitz

Phytologia (March 1991) 70(3):209-225.
TAXONOMIC NOTES ON CISTANTHE, CALANDRINIA, AND
TALINUM(PORTULACACEAE)
Mark A. Hershkovitz
Department of Botany, University of California. Davis, California 95616
U.S.A
ABSTRACT
New combinationsfor South American species ofCistanthearepro-
vided along with taxonomic notes. Also, a new section of Calandrinia
and anew combinationin Talinumare established. The nomenclaniral
changes reflect new morphologicaland cladistic data on Portulacaceae.
KEY WORDS: Argentina, Calandrinia. Chile, Cisianthe, Mono-
cosmia, Peru, Portalacaceae, Talinum. taxonomy
In a previous paper (Hershkovitz 1990a), Iestablished asectional taxonomy
and several combinations for Cistanthe Spach (Portulacaceae). Theseproposed
changes reflected corroboration by leaf morphological evidence (Hershkovitz
1990b) of Carolin's (1987) cladistic analysis of Portulacaceae, in which he
concluded that Cistanthe should be segregated from Calandrinia Kunth in
H.B.K. Ialso proposed (Hershkovitz 1990b) that MonocosmiaFenzl inEndl. cL'
Fenzl is phj'logenetically nestedin Calandrinia and should beincluded therein.
I provide below: (1) additional combinations for South American species of
Cistanthe, along with miscellaneous notes and lists of material examined: (2)
a sectional status for Monocosmia in Calandrinia] and (3) a new combination
in Talinum Adans. for a species currently classified in Calandrinia.
1. Cistanthe NEW COMBINATIONS AND NOTES
In my studies of leaf morphology in Portulacaceae (Hershkovitz 1990b), I
noted that specimenidentification andspecies delimitations ofSouth American
Cistanthe., heretofore classified in Calandrinia (of. Hershkovitz 1990a, 1990b)
and Philippiam,ra Kuntze (= Silvaea Philippi), were not adequately facilitated
209
PHYTOLOGJA
210 volume 70(3):209-225 March 1991
by existing floristic treatments. At the same time, I concurred with Carolin
(1987, in press) that Portulacaceae taxonomj' was poorly served by the reten-
tion ofthese species under their existing generic assignments. The first prob-
lem I circumvented by sampling leaves representative of the range of gross
morphological variation in Cistanthe as evident from herbarium specimens.
The second problem is more difficult to resolve because there is simply no
means short ofa detailed monographic study to determine which ofat least 60
combinations of Calandrinia and Philippiamra/Silvaea referable to Cistanthe
will prove to correspond to distinct species.
An exampleofthesort oftaxonomic problems Iencountered can be demon-
strated by comparing accounts by Reiche (1898) and Johnston (1929) of the
Cistanthe grandiflora(Lindlej^) Carolin ex Hershkovitz complex. According to
Reiche (1898), Cistanthe grandiflora represents a single polymorphic species
with a total of eleven synonyms variously referable to three more or less dis-
tinct varieties corresponding to Calandrinia grandiflora Lindley, Calandrinia
discolorSchrad., and Calandrinia speciosa Lehm. Johnston (1929, pp. 36, 37,
147), however, made several remarks indicating that he recognized at least
some ofthe variants as distinct species. In addition. Johnston (1929) charac-
terized two new species as members ofthe Cistanthe grandifloracomplex. The
difference in opinion cannot be attributed merely to a tendency by Johnston
to split taxa. For example, Reiche (1898) recognized four species in Cistanthe
sect. Philippiamra, while Johnston (1929, p. 39) wrote that " further studies
wiU show [three of] these species ...to be phases ofone variable species .. . ."
In the present paper, I provide new combinations for South American
species of Cistanthe examined in my leaf morphological studies, along with
remarks on possible synonyms and lists of examined specimens. This paper
brings the total number of recombined names for South American species to
17 - twelvefor Cistanthe sect. Cistanthe, three for C sect. Amarantoides, and
two for C. sect. Philippiamra (cf. also Hershkovitz 1990a). These figures wiD
increase eventually - I suspect additional research may well establish that the
actual numbersin each section are roughly twice these numbers. Unless other-
wise noted, the species recombined hereare presumed toinclude the synonyms
listed elsewhere, and no judgment is implicit in my failure to mention other
pertinent species of Calandrinia elsewhere recognized. There has never been
a comprehensive key to the South American species of Cistanthe. The most
extensive key, provided by Reiche (1898) for the Chilean species, emphasizes
variable vegetative and sometimes nebulous traits (cf. Carolin 1987). Given
the preliminary nature of the present work, however, I am unable to offer a
useful alternative. Identifications, unfortunately, will continue to require the
use ofthe keys in the cited floristic literature and simple description/specimen
comparisons.
For the present work, I consulted type descriptions and regional floristic
literature provided by Reiche (1898) and Philippi (1860a, 1860b, 1894a, 1894b)
Hershkoviiz: Taxonomic notes in PoTtula.ca.ceae 211
for Chile: Johnston (1928, 1929) for northern Chile and Peru; Macbride (1931,
1937) for Peru; Cullen (1953) for Argentina; and Peralta (1988) for the Ar-
gentinian province of Mendoza. I also considered descriptions of seed and
trichome morphology provided by Kelley (1973) for Cistanihe sections Cisian-
ihe and Amarantoides. and my own observations of herbarium material and
anatomical preparations thereof(Hershkovitz 1990b). I haveseen photographs
of supposed type material of many of the species (see below), but I generally
cannot confirm their authenticity.
A few years ago, Roger Carolin (SYD) provided me with an unpublished
manuscript that included several ofthe combinations listed below and several
others. Carolin proposed several combinations in Cistanihe that represent, I
believe, synonyms ofother species. In any case, I employ the "ez" notation to
indicate Carohn's authorship ofthe names contained in his manuscript.
A. Cistanihe sect. Cistanihe
Cistanihe sect. Cistanihe comprises all species of Calandrinia sect. Cistan-
ihe (Reiche), Calandnnia sect. Rosulaiat Reiche, Calandrinia seel. Arenanae
Reiche, and Calandrinia sect. Andinae Reiche (cf. Hershkovitz 1990b; KeUey
1973; Reiche 1897, 1898).
Cistanihe arenaria (Diels) Carolin eiHershkovitz, comb. nor. BASIONYM:
Calandnnia arenana Cham., Linnaea 6: 563. 1831.
SPECIMENS EXAMINED: CHILE. Aconcagua: Bertero 1348 (photo-
graph of type of Calandrinia chamissoi Barneoud in Gay, US ex G-DEL);
Morrison 16917 (CAS, NA). Atacama: Philippi s.n. (photograph of type
of Calandnnia polyclados Phihppi, US ex G-DEL). Coquimbo: Joseph 4507
(US); Joseph 4511 (US): Simon 441 (UC); Simon 454 (UC); Wagenknecht
18444 (F - two sheets, UC); Worth & Momson 16855 (CAS, NA); Zollner
6038 (L). MetropoHtana: Joseph 2788 (US). Nuble: Joseph 3990 (US).
Cisianthe arenaria is an annual with numerous branches spreading from
the base and rhombic to linear leaves. The species is apparently closely re-
lated to Cistanihe fenzlii (Barneoud in Gay) Carohn ex Hershkovitz, which is
perennial and has more erect branches (cf. Reiche, 1898). Calandrinia solisii
Philippi appears to be somewhat intermediate between Cistanihe arenariaand
Cistanihe fenzUi. so for now, I refer all intermediate plants to one species or
the other.
Cistanihe arenaria is widely distributed and highly variable (cf. Reiche,
1898). Some of the cited specimens (e.g., Wagenknecht 18444) correspond
perfectly to Reiche's (1898) key and description. The largest plants, however.
PHYTOLOGIA
212 volume 70(3):209-225 M^rch 1991
appear similar to supposed type material of Calandnnia polyclados Philippi,
which Reiche regarded as a synonym of Cistanihe fenzlii.
Joseph 4507 and 4511 key to and had been determined as Calandnnia
glaucopurpurea Reiche. Reiche's (1898) description of Calandnnia glaucopur-
purea^ however, also corresponds to that of Calandnnia chamissoi Barneoud
in Gay, which Reiche (1898) recognized as a variety of Cistantht arenana. Re-
iche (1898) recognized two other varieties of Cisianthe arenana, but I cannot
determine whether these are discrete.
Joseph 3990moreorless keys to and was determined as Calandnnia so/tsn,
but the specimen has a slender, rather than thick (perennial?; cf. Reiche 1897)
root. I am not convinced that the key trait (sterile inflorescence bracts present
vs. absent; Reiche 1898) and thedifferencein root morphology that distinguish
Calandnnia solisiifrom Cisianthe arenaria collectively warrant specific recog-
nition ofthe former.
I have elsewhere (Hershkovitz 1990b) cited Neger s.n. (M) and Zollner
10636{}\A) as Cisianthe arenana. The former specimen I believeis Cisianthe
fenzlii, although it also conceivably keys to Calandnnia solisii. The latter is
a mixed coDection with some arenana-]ike plants (see indets. listed below).
I also cited Zollner 9807incorrectly as Cisianthe glaucopurpurea (see indets.
listed below).
Cistanthe cephalophora (I.M. Johnston) Carolin ex Hershkovitz, comb,
not. BASION^'M: Calandnnia cephalophora I.M. Johnston, Contr.
Gray Herb. 85:35. 1929.
SPECIMENS EXAMINED: CHILE. Antofagasta: Pennell 13032 (US);
Werdermann 855 (CAS, UC, US).
This distinctive species has a capitate inflorescence and hairy seeds, but
otherwise resembles Cistantht longiscapa (Barneoud in Gay) Carolin ex Her-
shkovitz.
Cistanthe coquimbensls (Barneoud inGay) Carolin exHershkovitz, comb,
nov. BASIONYM: Calandnnia coquimbensisB&ineoud in Gaj', Fl. Chile
[Hist. Fisica y Politica de Chile, Botamca] 2: 483. 1847 ("1846").
SPECIMENS EXAMINED: CHILE. Locality not determined: Gay s.n.
(photograph of type, US ex P); Joseph 4973 (US). Coquimbo: Werdermann
881 {CAS, T,VC).
This species seems to be related to the Cisianthe longiscapa alliance {Ca-
landnnia sect. Rosulaiae Reiche, cf. Reiche 1898; see also below). The cited
specimens, both previously identified as C. coquimbensis, key to this species
more or less by default rather than by critical traits - the other species in
this section (Reiche's nos. 22-26) are quite distinct from Cistanihe coquimben-
sis. The cited specimens differ from Cistanihe longiscapa in: (1) having more
Hershkovitz: Taxonomic notes in Poriulacaceae 213
spreadingorprostrate,rather than erect,flowering branches; (2) havinglonger,
relatively narrower, and more distinctly petiolate leaves; and (3) having more
foliage leaves scattered along the flowering branches rather than just near the
base. In these respects, the specimens approach members of Calandriniasect.
Arenanae Reiche. The leaf venation pattern also appears distinct from that
of Cisianthe longiscapa (cf. Hershkovitz 1990b). The examined material has
ca. 15 stamens, rather than 6-8 as reported by Reiche (1898).
Cistanthe cymosa (Philippi) Hershkovitz, comb. nov. BASIONYM: Ca-
landrinia cymosa PhiHppi, Ann. Univ. Chile 85:192. 1893.
SPECIMENS EXAMINED: CHILE. Antofagasta: von Bayem s.n., "11.x.
1898" (M); Werdermann 85S{B, CAS, M, UC, US); Worth & Morrison 15816
(NA, UC). Atacama: Johnston .^757(US).
This species is related to the Cistanthe longiscapa alliance (cf. Reiche,
1898). Its distinctive feature is its nearh^ corymbiform inflorescence with nu-
merous congested flowers. The flowers are yellow and the seeds black and
shiny. The sepals in the cited specimens appear waxy to glutinous, a feature
noted for Calandrinia chrysantha I.M. Johnston, which Johnston (1929) re-
garded as most closely related to Cistanthe cymosa - they may prove to be
the same species.
Cistanthe fenzlii (Barneoud in Gay) Carolin ex Hershkovitz, comb. nov.
BASIONYM: CalandriniafenzliiBarneoud inGay, Fl. Chile [Hist. Fisica
y Poliiica de Chile, Botanica] 2:493. 1847 ("1846'").
SPECIMENS EXAMINED: CHILE. Locality not determined: Gay s.n.
(photograph of type, US ei P). Bjo-Bi'o: Joseph 4015 (US): Neger s.n., "Nov
1896" (two sheets, probably duplicates, M); Philippi s.n. (US).
In Reiche's (1898) key, Cisianthe fenzliiis set offfrom Cistanthe arenaria
only byvirtueofthemoreerect stemsin theformer. The description, however,
also notes that Cistanthe fenzlii has a thick root, and, indeed, the cited spec-
imens appear to be perennials. Judging from a Philippi-determined specimen
[Philippi 5.71.), Reiche (1898) was correct in regarding Calandrinia sangumea
Philippi as a synonym of Cistanthe fenzlii.
Reiche (1898) also listed Calandrinia polycladosPhilippi in synonymy with
Cistanthe fenzlii, but the photograph ofthe type ofthe former (see Cistanthe
arenaria above) is more highly branched, has broader leaves, and has less
congested inflorescences than the photograph ofthe type ofthe latter and the
specimens cited here.
Cistanthe grandifiora (Lindley) Carolin ex Hershkovitz, Phytologia 68:269.
1990.
PHYTOLOGIA
214 volume 70(3):209-225 March 1991
SPECIMENS EXAMINED: Cultivated: Peek 154 (NA). Without data:
one sheet (B); one sheet, no. 906253... S9(L). Without collector and locality:
993 (L). CHILE. Without locality: Bernhardt 33 (B). Aconcagua: Behn
5.71. (UC); Blood & Tremelhng 427 (NA); Buchtien s.n. (US); Eyerdam &
Beetle 24642{\}C)\ Hutchinson 176 (UC); Joseph 1764 (US); Morrison 16872
(NA, UC); Rose & Rose 19138 (US); Simon 446 (UC); Stebbms 8547 (UC);
Stebbins 8592 (UC); Werdermann 39 (UC); West 3959 (F, UC - two sheets,
US - two sheets); West 4574 (UC); Zollner 8364 (^-A)- Antofagasta: Worth
& Morrison 16133 (NA, UC). Atacama: Beetle 26143 (UC); Harding 22847
(US); Werdermann 405 (F, US); West 3868 (UC); Zollner 4079 (L). Bio-Bio:
Hutchinson 234 (UC, US). Coquimbo: Simon 447 (UC); Simon 462 (UC);
Simon 465 (UC); Wagenknechi 18562 (UC); Zollner 6327(L); Zollner 10284
(NA). Metropolitana: Hastings 268{\]C)\ Morrison 16816{\]C); Philippi s.n.
(US). O'Higgins: Aravena 33343 (UC).
I alluded above to the taxonomic problems in the Cistanthe grandiflora
complex. I am unable to sort these specimens into finer, discrete specific en-
tities. The Antofagasta and Atacama specimens, however, seem coarser and
more succulent than the rest. Three ofthese [Beetle 26143, West 3868, Worth
& Morrison 16133) were annotated as Calandrinia taltalensis I.M. Johnston
fide Johnston. I have elsewhere (Hershkovitz 1990b) cited West 3959 as Cis-
tanthe glauca (Schrad.) ined., and Worth & Morrison 16133b.s Cistanthe tal-
talensis (I.M. Johnston) med., but I chose not to recognize these combinations
at present.
Cistanthe lingulata (Ruiz Lopez & Pavon) Hershkovitz, comb. nov. BA-
SIONYM: Talinum lingulatum Ruiz Lopez I: Pavon, Syst. Veg. Fl. Pe-
ruv. Chil. 115. 1798.
SPECIMENS EXAMINED: PERU. Ancash: Ferreyra 13532 (US). La
Libertad: Lopez Miranda 57.^ (US). Lima: Domiey(numberillegible, possibly
170: P); Dombey (number illegible, possibly 716; P); Ferreyra IO486 (US);
Ruiz Lopez & Pavon s.n. (photograph of type, US ex G-DEL); Weberbauer
7432 (US).
According to Macbride (1937), Cistanthe lingulata differs from Calandrinia
quivensis Macbride only in having a less exserted capsule. He noted, however,
that capsule exsertion may reflect capsule maturity, so that the two species
might be the same. Therefore, I have included all material keying to either
speciesin Cistanthe lingulata. Ofthe cited specimens, the most extremeexser-
tion of the capsule occurs in both of the Ferreyra specimens.
Cistanthe lingulata also apparently intergrades with Cistanthe weberbaueri
(Diels) Carolin ei Hershkovitz. According to Macbride (1937), Cistanthe we-
C
berbaueri differs from lingulata in having an "umbellate-fasciculate" rather
than "racemose or paniculate" inflorescence. Such a differencetypifies younger
Hershkovitz: Taxonomic notes in Portulacaceae 215
and more mature inflorescences, respectively, in Cisianthe. My determinations
emphasize apparent distinctions in leaf morphology: The leaves of specimens
referable to Cistanthe lingulaia are genercJly narrower (<10 mm) than those of
Cisianthe weberbauen (>10 mm), and, while their venation patterns are sim-
ilar, the vein orders of Cistanthe weberbauen are more distinct and the vein
reticulum appears denser even in comparably sized leaves (cf. Hershkovitz
1990b). \^'eberbauer 74S2. however, was cited as Calandnnia weberbauen by
mm
Macbride (1937) but has linear-filiform leaves less than 5 wide. The in-
florescence is immature in this specimen, so I cannot evaluate the other kej-
characters.
Cistanthe longiscapa (Barneoud in Gay) Carolin ex Hershkovitz, comb,
nov. BASIONYM: Calandnnia longiscapa Barneoud in Gay, Fl. Chile
[Hist. Fisica y Politico de Chile, Botamca] 2:492. 1847 ("1846").
SPECIMENS EXAMINED: CHILE. Without locality: Gay s.n. (photo-
graph of type of Calandnnia longifolia [sic], US ex P). Antofagasta: Werder-
mann IO46 (leg. Francke; UC, US). Atacama: Johnston 5034 (US); Morong
1267{KA, US); Philippi s.n. (US); Werdermann 445 (F, UC); Worth & Mor-
nson 16164 (NA, UC); Worth & Momson 16267{SA, UC); Zollner 4083(L);
Zollner 9291 (NA); Zollner 11917{KA).
Cisianthe longiscapa is distinguished from C. cephalophora and C cymosa
by its more open inflorescence. Vegetatively, the.se three species are similar in
having relatively broad and tending towards orbicular leaves crowded near the
base of the plant and typically a pair of subopposite leaves at the base ofthe
inflorescence.
I tentatively refer Calandrima hioralis Philippi to Cistanthe longiscapa.
From the examined material, I cannot identify the putative differences be-
tween these species in bract morphology (cf. Reiche 1898). Pubescent seeds
supposedly characterize Cisianthe longiscapa, while glabrous seeds character-
ize Calandnnia litoralis (Kelley 1973; Reiche 1898). The specimens cited
above have variously pusticulate-tomentose or glabrous seeds, but are other-
wise very similar. According to Reiche (1898), Cistanthe longiscapa flowers in
September-October, while Calandrinia litoralis flowers in December-January.
Most of the cited specimens have September-October collection dates, while
Johnston 5034 lists November 25, and Zollner 408S]ists January 12. The last
two specimens are clearly in more mature stages than the September-October
collections.
The Atacama collections are mostly from the vicinity of Copiapo. The
Antofagasta collection {Werdermann IO46) is unusual and may be misdeter-
mined. This specimen includes a fragment of an inflorescence and a w^hole
plant that appears to have been prostrate with a flowering scape barely 5 cm
long. The bracts and sepals are not conspicuously etched with black markings.
The absence of these etchings is unusucd in Cistanthe sect. Cistanthe.
PHYTOLOGIA
216 volume 70(3):209-225 March 1991
Cistanthe macrocalyx (Hauman) Carolin ex Hershkovitz, comb. nov. BA-
SIONYM: Calandnnia macrocalyxEa-uman, Ann. Soc. Cient. Argentina
86:253. 1919.
SPECIMEN EXAMINED: ARGENTINA. San Juan: Casiellanos 15587
(US).
Cullen (1953) and Peralta (1988) recognized this species as distinct from
Cisianthe picia(Gillies exArn. in Cheek) Carolin txHershkovitzin their treat-
ments for Argentina and Mendoza, respectively. It is distinguished from the
latter by its smaller, oblanceolate leaves and larger, relatively broader calyx.
The specimen cited above agrees with the description in leafmorphology, but
the flowers have largely been lost.
Cistanthe panlculata (Ruiz Lopez k Pavon) Carolin txHershkovitz, comb.
nov. BASIONYM: Calandnnia paniculaia Ruiz Lopez k Pavon, Syst.
Veg. Fl. Peruv. Chil. 114. 1798.
SPECIMENS EXAMINED: PERU. Ancash: Dillon, ei al. 4627{F); Dil-
lon, et al. 46S3{F)] Dillon & Santisiebau 4706 {F)] Ferreyra 6350 (US); Fer-
reyra 11576{VS); Ferreyra 12022{VS)] Ferreyra 12611 (US); Hutchinson 1292
(US, UC); Stork, et al. 9169 (UC); Weberbauer I484 (B). Arequipa: Dillon,
et al. 3265 (F); Dillon, et al. 3292 (F); Dillon & Dillon 3843 (F); Worth &
Morrison 15761 (NA, UC). La Libertad: Dillon, ei al. 2718(F); Dillon, et al.
4656 (F); Dillon, et al. 4693 (F). Lima: Soukup 2156 (US); Stork & Vargas
9363{VC). Tacna: Dillon, et al. 3384 (F); Dillon & Dillon 3662{F); Ferreyra
12502 {VS); Ferreyra 12507{VS); Ferreyra 12519 (VS).
I tentatively regard Cistanthe paniculata and Calandnnia ruizii Macbride
as belonging to a single polymorphic species based on hints to this effect in
the literature and my inability to readily sort the specimens. According to
Macbride's (1937) key and description, Cistanthe paniculata differs from Ca-
landnnia ruizii in having a more exserted capsule, more densely pubescent
seeds,and "obovate-oblong" ratherthan "broadly obovateorsubrotund'" leaves
with acuminate rather than "merely apiculate" apices. Macbride (1937) listed
Calandnnia adcnosperma I.M. Johnston as a synonym of Cistanthe panicu-
lata, but he had previously noted (Macbride 1931) that Calandrinia ruizii
was also similar to Calandnnia adenosperma. Johnston's (1928) description
of Calandnnia adenosperma essentially bridges the gap between the descrip-
tions of Cistanthe paniculaia and Calandrinia ruizii: As in Calandnnia ruizii,
the sepal and capsule lengths in Calandrinia adenosperma are equal, and the
upper leaves are subamplexicaul and orbicular-ovate with obtuse apices.
Cistanthe paniculata, Calandrinia ruizii, and Calandrinia adenosperma all
reportedly havehairy seeds (Johnston 1928; Kelley 1973; Macbride 1937), but
I have found glabrous seeds, apparently not correlated with any other trait in
Hershkovitz: Taxonomic notes in Portulacaceae 217
several of the specimens. Label data from the cited collections also indicate
that flower color (yellow vs. red) also varies independently ofother traits (cf.
Macbride 1937).
Macbride (1931, 1937) and Johnston (1928) both believed Cistanthe pan-
iculaia to be especially closely related to Cistanthe grandiflora. Preliminary
studies (Hershkovitz 1990b) indicate that thelatter, alsoapolymorphic species
complex, is distinctive in terms of leaf venation and perhaps also stomatal
morphology and stomatal density. In leafvenation, Cistanthe paniculatamore
closely resembles Cistanthe longiscapaiha.n Cistanthe /arnprosperma(I.M.John-
ston), ined. (Hershkovitz 1990b).
I have elsewhere (Hershkovitz 1990b) cited Ferreyra 10J,86a.s Cistanthe cf.
ruizn (Macbride) ined.
Cistanthe picta (Gillies ex Arn. in Cheek) Carolin ex Hershkovitz, comb,
nov. BASIONYM: Calandnnia picta Gillies ex Arn. in Cheek, Edin-
burgh J. Nat. Geogr. Sci. 3:355. 1831.
SPECIMENS EXAMINED: ARGENTINA. Mendoza: Gerlh ]07{L). San
Juan: Castellanos 15588 (US).
CHILE. Without locality: Gay s.n. (photograph oftypeof Calandnnia ob-
longijolia Barneoud in Gay, US ex P). Without province: Meyen s.n. "Cord.
St. Fernando" (B). Aconcagua: Hutchinson 98 (UC, US); Morrison 17304
(NA, UC); West 5169 (UC); ZoUner S642 (L); Zollner 6548 (NA); Zollner
11032{KA). Antofagasta: Werdermann 1038 {VS,VC). Atacama: Johnston
6218 (US). Coquimbo: Worth & Morrison 16584 (NA, UC). Metropolitana:
Grandjoi s.n., "XII. 32^" (M); Grandjots.n., "XII. 34'" (M): Hastings 411 (UC);
Joseph 2P55(US); Kuntze s.n., "I 92" (US); Morrison 16786{'^A, UC); Wer-
dermann 648[VC, CAS). O'Higgins: Gay s.n. (photograph, not type material
fide Peralta [1988], US ex G-DEL); Pennell 12279 (F).
Cistanthe picta and related members ofReiche's (1897, 1898) Calandnnia
sect. Andinae are readily distinguishable from other Cistanthe by their: Thick,
perennial roots; spreading habit; short stems bearing numerous, usually obo-
vate to ovate, petiolate leaves; and congested cymules occurring immediately
above the leafy stems and bearing relatively short pedicelled flowers and per-
sistent, conspicuously black etched bracts. Also, the plants occur at higher
elevation (mostly 2500 to 3500 m) than other species of Cistanthe. Despite
the distinctiveness of this group, Reiche's (1898) key separates Calandnnia
sect. Andinae from other members of Cistanthe sect. Cistanthe only with re-
spect toleafshape, although leafshape is highly variable throughout thegenus
(Hershkovitz 1990b).
Cistanthe picta appears to represent a highly polymorphic entity from
which notable variants have been segregated. Without seeing additional and
type material, I cannot make any judgment regarding the distinctiveness of
PHYTOLOGIA
218 volume 70(3):209-225 March 1991
most of the segregates recognized or listed in synonymy by Reiche (If
Those recognized here are C. piciavar.frigida (Barneoud in Gay) Hershkovitz
(see below) and Cistanthe macrocalyx (see above). I do note, however, that
Johnston annotated his Atacama collection (J. 6218) as possibly representing
mm
a new species. It has conspicuous reniform bracts ca. 5 wide - roughly
twice the norm.
A surprisingly large number of specimens of Cisianthe picia have been
annotated, sometimes emphaticeilly, mostly anonymously, as Calandrinia are-
naria Cham. The type and subsequent descriptions and illustrations of Calan-
drinia arenaria Cham. (Chamisso 1831; Hooker i: Arnott 1833: Lindley 1833)
all clearly pertain to Cistantht artnana as delimited by Reiche (1898).
Cistanthe picta (Gillies ex Arn. in Cheek) Carolin ex Hershkovitz var.
frigida (Barneoud m Gay) Hershkovitz, comb. nov. BASIONYM: Ca-
landrinia frigida Barneoud in Gay, Fl. Chil. [Hist. Fisica y Politica de
Chilt, Boiamca] 2:495. 1847 ("1846").
SPECIMENS EXAMINED: CHILE. Without province: Gay s.n., "1333
34'" (P). Coquimbo: Gay s.n. (photograph oftype of Calandriniafrigida. fide
Peralta [1988], US ex P): Momson 16992 (NA, UC).
Peralta (1988) listed this variety in synonymy with the species. This vari-
ety occurs at higher altitudes than the rest of the species, which may account
for its reduced morphology. Kelley (1973), however, reported glabrous, coUic-
ulate seeds, like those of Momson 16992., in Cisianthe picia var. frigida. The
Gay collection has no seeds. Leaves of this specimen also lack sinuous and
ribbonlike veins that occur in Cisianiht picia and nearly all other species of
Cisianthe (Hershkovitz 1990b).
Cistanthe weberbaueri (Diels) Carolin ex Hershkovitz, comb. nov. BA-
SIONYM: Calandrinia weberbaueri Diels, Bot. Jahrb. Syst. 37:399.
1906.
SPECIMENS EXAMINED: PERU. Arequipa: Ferreyra 11723 {VS); Fer-
reyra 12006 (VS). Lima: Weberbauer 5321 (F).
As noted above, the distinction between Cisianthe weberbaueriand C. Im-
gulaia is tenuous. Weberbauer 5321 is unusual: It has the broadest leaves
of any specimens I have examined in the lingulaia/weberbaueri complex, and
it has distinctly hairy seeds. The seed morphology of the other cited spec-
imens accords with Kelley (1973), who reported pusticulate-tomentose seeds
in Cistanthe lingulaia, Cisianthe weberbaueri. and Calandrinia quivensis. We-
berbauer 5321 was apparently miscited by Macbride (1937) as "5221."
Specimens not determined: The specimens listed below belong to species
of Cisianthe sect. Cistanthe listed above, species not yet transferred from Ca-
landrinia., or possibly to new species. The specimens are all from Chile.